INTERNATIONAL JOURNAL OF HUMAN-COMPUTER INTERACTION, 77(3). 427-144Copyright © 2004, Lawrence Eribaum Associates, Inc.

Human Body as the Medium in Dance Movement

Mamiko SakataFukushima College

Mariko ShibaFaculty of Human Development, Kobe University

Kiyoshi MaiyaResearch Institute for Higher Education, Kobe University

Makoto TadenumaThe Science and Technical Research Laboratories of NHK

To clarify mechanisms involved in the human recognition of body-mediated informa-tion, this study examines what kinds of impressions and emotions are perceived frombody movements in dance. The study attempted to develop multiple regression mod-els to investigate the effects of physical movement factors on impressions and emo-tions. The regression analyses show that people perceive emotional information fromthe body-mediated information, even without facial expressions. Those results sug-gest that the body does not only have an additional or secondary function, but alsoplays an important role as independent media in human communications. Emotionsare actually expressed and impressed, that is, communicated through the humanbody as the medium.

1. INTRODUCTION

In interpersonal communication, people need to not only know each other's verballanguage, but also have a skill to read nonverbal information that is reflected inpeople's breathing manner, tone of voice, facial expression, posture, and gestures

This study has been funded as part of the various projects such as "Verification of Basic EmotionsCommon to Different Emotionai Expressions and Comparison Between Different Cultures" of Funda-mental Research in Grant-in-Aid for Scientific Research, and "Basic Study on Correlation between Dif-ferent Emotional Media" consigned by the ATR Media Integration and Communication Research Labo-ratories (1998-2000) and the project "Study on Emotional Information in Physical Expression," in whichwe were research assistants in 2000. We acknowledge Saori Hanaki, a contemporary dancer, for her as-sistance on this project.

Requests for reprints should be sent to Manuko Sakata, Fukushima College, 1-1 Chigoike,Miyashiro, Fukushima City, 960-0181 JAPAN. E-mail: MamitamaM@aol.com

428 Sakata et al.

on a real-time basis (Yoshida, Renge, Kagawa, Sako, & Maiya, 1989). Such nonver-bal information can be called "information conveyed through human body as themedium" because emotional information is transmitted by using our own body asthe media for communication. This kind of body-mediated information sometimescontains more messages than audio information such as language (Sakata, Shida, &Iwadate, 2000).

Wundt (1973), the founder of modern psychology, stressed the importance ofgestures as a means of expressing oneself. He further stated that unintentional,more impetuous emotions are expressed before intentional communication. Inother words, gestures that unconsciously appear are considered among the mostprimitive forms of information conveyed through human body as the medium, di-rectly connected to human emotions.

To achieve a natural interface, nonverbal interface incorporating informationconveyed through human body as the medium is indispensable. Brisk research isbeing done on bodily interfaces that convey gestures, postures, and the like(Kurokawa, 1992; Kurokawa, Morich, & Watanabe, 1989). The central themes of theresearch are gesture language and sign language with their clear-cut grammar andclause structure used for conveying the intentional messages (Imagawa et al., 2000;Lim & Tamura, 1999). However, much of the body-mediated information conveyedin human communications have neither grammar nor clause structure. For actualcommunication to take place, it is necessary to clarify the mechanism of body-me-diated information directly linked to human emotions. For that purpose, it is cru-cial to build up a relational model concerning body movements and emotional ex-pressions, whose correlation is still ambiguous.

We have conducted a basic research to clarify mechanisms involved in the hu-man recognition of body-mediated information, and in particular, we focus on therelation between impressions and emotions, conveyed through body-mediated in-formation, and body movements expressing the body-mediated information. Outof body-mediated information, dance is a form in which human emotional imagesare richly expressed. One can say that both dance, which is conscious and inten-tional body expressions by humans, and facial expressions and gestures, which areunconscious and unintentional body expressions, are forms of body-mediated in-formation in a sense that the human body acts as the medium.

Regarding the connection between dance movements and emotions,Matsumoto (1985) proposed a hypothesis in which "there are some potentiallycommon characteristics at the base of uniqueness in expression." They continue re-search on the "emotional value of movements" to verify the hypothesis and as partof efforts to scientifically clarify the structure of expression in dance. Thus,Matsumoto's theory is widely applicable to body-mediated information, withoutjust being confined to dance, and we are working to clarify the structure and regu-larity present in the body-mediated information based on the theory—Matsumoto's 7 Motives (Sakata et al., 2000; Sakata, Uchida, Hama, Shirakawa, &Maeda, 2001).

Matsumoto's 7 Motives (1985, see Figure 1) empirically verified the connectionbetween images and movements. The 7 Motives are the seven dance movementsthat Matsumoto set up to be the minimum unit of expressions, in which each mo-tive contains clearly distinguishable emotional value. Matsumoto added the arms

M.M. ; =120

1 tount=2 steps. 3 tount=l turnEnergy HighDesign Skip TXirn

Happy Mv.

M.M. J =100Time

4 count=l phraseEnergy HighDesign Straight

Sharp Mv.

Time M.M. J =36

4 count-i stepEnergy LowDesign Up&nowii,ABymmetnea

Mv.

Time M.M. J =96 ±Even Rhythm 4

JJJJJJJJ icount=li?Knergy NormalDesign Symmetrical, Balance

t=lstep

Natural Mv.

Time M.M. ; =15

JJJ3 stepa^l phrast:

Energy LowDesign Symmetrical, Balance

Solemn Mv.

Time M.M. J=192

• JJ.JJ-J10 count=l phrase

Energy HighDesign Jump

Dynamic Mv.

Time M.M. J =92

JJJJJJ1 count=l step, 3 count=l turnEnergy NormalDesign Smooth-faced turn

Flowing Mv.

FIGURE 1 Matsumoto's 7 Motives. Mv. - motive.429

430 Sakata et al.

raising and lowering and closing and opening n:iovements to the ordinary walkingmovement and then set up the movements that involve the entire body, includingexhalation and inhalation, as the "basic pattern" and called it the Natural Mv.Based on the Natural Mv., she varied the time, energy, and design^ factors of move-ments, while retaining the characteristics of the Natural Mv. to the extent possibleand created three movements at both extremes. She named them as Happy Mv.,Lonely Mv., Sharp Mv., Solemn Mv., Dynamic Mv, and Flowing Mv. Mv. is the ab-breviation of "Motive" used by Matsumoto.

This article focuses on two points. One is to clarify the kinds of impressions andemotions perceived from body movements in dance using Matsumoto's 7 Motives.Another is to develop relation models to investigate the effect of physical move-ment factors on impressions and emotions conveyed through the body-mediatedinformation.

2. METHOD

2.1 Raters and Dates of Experiments

Thirty-eight raters (12 men, 26 women, with an average ageof 20.6, SD -1.36) partici-pated in this experiment. The raters had no previous dance experience. The experi-ment was conducted in January 2000 in separate classrooms. To allow all the raters tosufficiently observe the stimuli, dance images were projected on a large screen.

2.2 Displayed Stimuius

Matsumoto's 7 Motives were newly acted out by a dancer and recorded on video-tape, which was used as the displayed stimulus (see Figure 1). Because the mainobjective of this study is to clarify the relation between body movements in danceand the impressions perceived from such movements, we used dance movementsshown by one dancer capable of differential representation of the 7 Motives as thedisplayed stimuli.

Dancer and rehearsal. The performer (dancer) was a 22-year-old female stu-dent n^ajoring in dance with 11 years of training in modern dance, classical ballet,and contemporary ballet. Before starting the filming, she was briefed on the experi-ment, and a 2-week rehearsal followed. During the rehearsal, the authors and twoother experts familiar with the 7 Motives were present, and the performer re-hearsed until she was able to represent the 7 Motives in a loyal manner, withoutchanging her facial expressions.

e, energy, and design are the factors of a physical mtwement. They are correlated, so when anyof these factors change, the overall representation will change. Time factor works on the time aspect ofmovement, energy, on the energy aspect of movement, and design, on the morphological aspect ofmovement (Matsumoto, 1968). in formulating the 7 Motives, Matsumoto changed these factors to giveeach motive a unique image.

Human Body as the Medium 431

FUming condition. The performance of the 7 Motives was filmed with a digi-tal video camera, under the conditions shown in Figure 2. On the day of filming,the performer rehearsed for about 30 min and then performed each motive threetimes. When filming, care was taken to include the entire body in the visual frameof the camera.

Stimuli selection. On examining the video, the two dance researchers se-lected one performance showing the best representation for each motive. The sevenrecordings selected were patched together in random order, and were named Mo-tive 1 through Motive 7. Table 1 shows the order of motives and the duration of per-formance of the motives, which were set according to Matsumoto (1985). Thelength of performance was the minimum needed for the motive to capture its im-age (Matsumoto, 1985).

2.3 List of Rating Terms

In this experiment, rating terms were used from Matsumoto's Check List I andCheck List II, which were selected after a series of Matsumoto's research on danceterminology and the emotional values of physical movements, along with Ekman'sbasic emotions.

1.6mrange(5m x 5m)

0.95

perform 17m camera

FIGURE 2 Filming Condition.

Table 1; Order and Time of the Stimuli

Order of Display

Motive 1Motive 2Motive 3Motive 4Motive 5Motive 6Motive 7

Matsumoto's 7 Motives

Happy MotiveLonely MotiveSharp MotiveNatural MotiveSolemn MotiveDynamic MotiveFlowing Motive

Duration

8 sec27 sec4sec11 sec34 sec4sec8sec;

432 Sakata et al.

Check List I. Matsumoto's Check List 1 (1985) contains a random arrange-ment of 18 opposing terms that she selected in terms of time, energy, and design.Table 2 shows the three categories and 18 terms in Matsumoto's Check List 1.

Check List ii. Matsumoto's Check List II (1985) is a random arrangement of42 tern^s in seven categories (6 terms in each category) that she edited out of thelarge number of adjectives and adverbs, which had been categorized into synony-mous and related terms. Table 3 shows the seven categories and the 42 terms inMatsumoto's Check List II.

Basic emotions. This study used the six basic emotions postulated byEkman and Friesen (1975), which are happiness, loneliness, surprise, fear, anger,and disgust, as rating terms.

2.4 Procedure

After the raters were asked to glance over the list of rating terms, they were shownthe display stimulus (motive video) showing the motives one by one. After thesame motive was shown twice, the vision changed to an interval. During the inter-val, the raters were asked to fill in the answer sheets with the following questions.Before being shown the stimulus, the raters were instructed to "rate the perfor-

Table 2: Matsumoto's Check List I

Time speedy—slowaccentuated—smoothirregular—regular

Energy strong—weakheavy—lightvariable—constant

Design linear—curvaceousexpansive—diminutivebalanced—unbalanced

Table 3: Matsumoto's Check List II

happy, bright, light-hearted, merry, exciting, humorousflowing, tender, soft, elegant, graceful, warmlonely, dark, viscous, quiet, sad, weaknatural, commonplace, ordinary, unaffected, calm, simplesolemn, majestic, imposing, stable, holy, deepsharp, cold, mechanical, rigid, threatening, aggressivedynamic, powerful, large, gleeful, heroic, lively

Note. The underlined word is the most representative in each category

Human Body as the Medium 433

mance based only on body movements since the performer would not show any fa-cial expressions."

1. What impressions they received from each motive (mandatory selection ofthree terms from Check List II).

2. What kind of motion each motive showed (mandatory selection of threeterms from Check List I).

3. How much of the six basic emotions (happiness, surprise, loneliness, fear,anger, disgust} was expressed in each motive, rated in four-point scalesranging from 1 {not expressed) through 4 {very clearly expressed).

The experirr\ent was conducted during the class period. Video screeningtemporarily stopped during the interval and was resumed after ensuring that allraters had written their answers. The duration of the interval was about 1 min.

3. RESULTS

For the collection and analysis of the data, the statistics software SPSS Base 10.OJ forWindows was used.

3.1 Rating on Impressions of Motives

Table 4 shows the results of mandatory selection of rating terms from Check List 11.From Table 4, the most frequently selected three terms are light-hearted, happy,

and gleeful for Motive 1 (Matsumoto's Happy Mv.); deep, quiet, and lively for Motive2 (Matsumoto's Lonely Mv.); sharp, mechanical and powerful for Motive 3(Matsumoto's Sharp Mv.); quiet, soft, and calm for Motive 4 (Matsumoto's NaturalMv.); imposing, quiet, and solemn for Motive 5 (Matsumoto's Solemn Mv.);light-hearted, dynamic, and happy for Motive 6 (Matsumoto's Dynamic Mv.); andflowing, elegant, and soft for Motive 7 (Matsumoto's Flowing Mv.). From these re-sults, more than three tern:\s were selected by more than 30% of the raters for Mo-tive 1, Motive 3, Motive 6, and Motive 7. As for Motive 2 and Motive 5, one termeach was selected by over 307o of the raters. For all the motives except Motive 4, theresults suggest that the raters had perceived emotions particular to each motive.The terms listed here almost perfectly match the terms that Matsumoto selected foreach motive category, except for Motive 4 (see Table 3). It is assumed that the raterswere able to read the images contained in each motive, except for Motive 4.

3.2 Rating on Motions of Motives

Table 5 shows the results of mandatory selection of three terms from Check List I tocharacterize each stimulus.

434 Sakata et al.

Table 4: Result of Check List II

Light-heartedBrightHappyMerryExcitingHumorous

StiftTenderFlowingElegantGracefulWarm

SadDarkViscousQuietLonelyWeak

NaturalCommonplaceOrdinaryUnaffectedCalmSimple

StableMajesticImposingSolemnHolyDeep

RigidColdMechanicalSharpThreateningAfigressive

DynamicPowerfulLargeGleefulHeroicLively

Motive 1

24 (63.2)4(10.5)

13(34.2)1 (2.6)0 (0.0)3 (7.9)

12(31.6)4 (10.5)

12(31.6)4 (10.5)2 (5.3)7 (18.4)

0 (0.0)0 (0.0)0 (0.0)1 (2.6)0 (0.0)0 (0.0)

0 (0.0)0 (0.0)0 (0.0)1 (2.6)0 (0.0)0 (0.0)

1 (2.6)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)

0 (0.0)0 (0.0)0 (0.0)1 (2.6)0 (0.0)0 (0.0)

6(15.8)0 (0.0)1 (2.6)

13(34.2)0 (0.0)5 (13.2)

Motive 2

0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)

3 (7.9)1 (2.6)0 (0.0)0 (0.0)0 (0.0)2 (5.3)

9 (23.7)9 (23.7)6(15.8)

10 (26.3)7 (18.4)2 (5.3)

0 (0.0)0 (0.0)0 (0.0)6 (15.8)4 (10.5)2 (5.3)

3 (7.9)1 (2.6)6 (15.8)4 (10.5)3 (7.9)

19 (50.0)

4(10.5)2 (5.3)1 (2.6)0 (0.0)0 (0.0)0 (0.0)

0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)

10 (26.3)

Motii'e S

2 (5.3)1 (2.6)0 (0.0)0 (0.0)1 (2.6)3 (7.9)

0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)

0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)

0 (0.0)0 (0.0)1 (2.6)0 (0.0)0 (0.0)7(18.4)

2 (5.3)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)

11 (28.9)3 (7.9)

23 (60.5)24 (63.2)

8(21.1)11 (28.9)

0 (0.0)12(31.6)

2 (5.3)0 (0.0)3 (7.9)0 (0-0)

Motive 4

0 (0.0)1 (2.6)0 (0.0)1 (2.6)0 (0.0)0 (0.0)

10 (26.3)5 (13.2)5 (13.2)2 (5.3)0 (0.0)1 (2.6)

2 (5.3)1 (2.6)0 (0.0)

11 (28.9)4 (10.5)4(10.5)

7(18.4)4(10.5)6(15.8)5 (13.2)9 (23.7)7(18.4)

8 (21.1)0 (0.0)0 (0.0)4 (10.5)7 (18.4)0 (0.0)

0 (0.0)2 (5.3)2 (5.3)0 (0.0)1 (2.6)0 (0.0)

1 (2.6)0 (0.0)1 (2.6)0 (0.0)1 (2.6)2 (5.3)

Motive 5

0 (0.0)0 (0.0)0 (0.0)0 (0.0)2 (5.3)0 (0.0)

2 (5.3)0 (0.0)2 (5.3)1 (2.6)0 (0.0)2 (5.3)

4 (10.5)6 (15.8)5 (13.2)8(21.1)3 (7.9)1 (2.6)

1 (2.6)1 (2.6)0 (0.0)0 (0.0)5 (13.2)2 (5.3)

3 (7.9)6 (15.8)

13 (34.2)8(21.1)7 (18.4)7 (18.4)

1 (2.6)2 (5.3)2 (5.3)1 (2.6)0 (0.0)1 (2.6)

1 (2.6)1 (2.6)7 (18.4)0 (0.0)3 (7.9)6 (15.8)

Motive 6

26 (68.4)12(31.6)13(34.2)

7(18.4)0 (0.0)2 (5.3)

2 (5.3)0 (0.0)4(10.5)1 (2.6)4(10.5)2 (5.3)

0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)

0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)

1 (2.6)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)

0 (0.0)0 (0.0)1 (2.6)0 (0.0)0 (0.0)0 (0.0)

14 (36.8)0 (0.0)3 (7.9)

10 (26.3)0 (0.0)

12(31.6)

Motii?e 7

4(10.5)4(10.5)5 (13.2)0 (0.0)0 (0.0)0 (0.0)

15 (39.5)5 (13.2)

22 (57.9)16 (42.1)10 (26.3)

3 (7.9)

0 (0.0)0 (0.0)0 (0.0)1 (2.6)0 (0.0)0 (0.0)

1 (2.6)1 (2.6)0 (0.0)2 (5.3)7 (18.4)0 (0.0)

3 (7.9)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)

0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)

3 (7.9)0 (0.0)4 (10.5)3 (7.9)0 (0.0)4(10.5)

Note. Percentages are given in parenthesis.

From Table 5, there was a tendency for selecting plural terms in the time category,and one term was selected from the energy category for each motive. Therefore, wecan assume that the raters were able to perceive the characteristics of each movementfor every motive in terms of time and energy aspects, and also that Motive 1, Motive3, and Motive 7 tended to conjure up spatial movement images in the raters.

Human Body as the Medium 435

Table 5: Result of Check List I

TimeSpeedyAccentuatedIrregularStowSmoothRegular

EnergyStrongHeavyVariableWeakUghtConstant

DesignLinearExpansiveBalancedCurvaceousDiminutiveUnbalanced

Motive J

4(10.5)12(31.6)0 (0.0)5 (13.2)

26 (68.4)9 (23.7)

2 (5.3)0 (0.0)0 (0.0)1 (2.6)

26 (68.4)2 (5.3)

0 (0.0)0 (0.0)5 (13.2)

20 (52.6)0 (0.0)0 (0.0)

Motive 2

0 (0.0)0 (0.0)1 (2.6)

34 (89.5)15 (39.5)5 (13.2)

5(13.2)22 (57.9)

0 (0.0)9 (23.7)0 (0.0)4(10.5)

1 (2.6)2 (5.3)2 (5.3)8(21.1)4(10.5)2 (5.3)

Motive 3

21 (55.3)16 (42.1)

1 (2.6)0 (0.0)0 (0.0)

16(42.1)

13(34.2)2 (5.3)9 (23.7)0 (0.0)1 (2.6)1 (2.6)

29 (76.3)2 (5.3)2 (5.3)1 (2,6)0 (0.0)0 (0.0)

Motive 4

0 (0.0)0 (0.0)0 (0.0)

27(71.1)16(42.1)15 (39.5)

1 (2.6)2 (5.3)0 (0.0)

14(36.8)3(7.9)

10 (26.3)

5 (13.2)5 (13.2)6(15.8)9 (23.7)0 (0.0)1 (2.6)

Motiz'e 5

0 (0.0)1 (2.6)1 (2.6)

33 (86.8)11 (28.9)

1 (2.6)

10 (26.3)23 (60.5)

1 (2.6)3 (7.9)1 (2.6)5 (13.2)

3 (7.9)7(18.4)3 (7.9)8(21.1)1 (2.6)1 (2.6)

Motive 6

29 (76.3)18 (47.4)1 (2.6)0 (0.0)4 (10.5)2 (5.3)

3 (7.9)0 (0.0)3 (7.9)0 (0.0)

27(71.1)2 (5.3)

2 (5.3)9 (23.7)6 (15.8)4(10.5)0 (0.0)1 (2-6)

Motive 7

1 (2.6)1 (2.6)1 (2.6)

19 (50.0)34 (89.5)3 (7.9)

1 (2.6)1 (2.6)1 (2.6)0 (0.0)

15 (39.5)6(15.8)

0 (0.0)3 (7.9)6(15.8)

21 (55.3)0 (0.0)0 (0.0)

Note. Percentages are given in parenthesis.

3.3 Rating on Selection of Emotionai Terms

The rating results using the 4-level extreme scales for the six basic emotions (happi-ness, surprise, loneliness, fear, anger, and disgust) ranged from 1 {not expressed), 2{not expressed weU), 3 {expressed well), to 4 {very dearly expressed). The average pointand standard deviation were obtained and tabulated in Table 6.

From Table 6, happiness scored the highest point for Motive 1 {Matsumoto'sHappy Mv.); sad for Motive 2 (Matsumoto's Lonely Mv.); anger and surprise forMotive 3 (Matsumoto's Sharp Mv.); happiness and sad for Motive 4 (Matsumoto'sNatural Mv.); sad for Motive 5 (Matsumoto's Solen^n Mv.); happiness and surprisefor Motive 6 (Matsumoto's Dynamic Mv.); and happiness for Motive 7

Table 6: Evaluated Scores and Standard Deviations of Basic Emotions

HappinessSurpriseLonelinessFearSurpriseDisgust

Motii'e 1

Score

3.74"*1.971.391.111.051.05

SD

0.600.720.600.310.230.23

Motive 2

Score

1.791.323.16""1.951.551.68

SD

0.910.620.920.870.860.%

Motive 3

Score

1.372.50"1.451.762.84™2.08

SD

0.491.060.600.851.131.00

Motive 4

Score

2.37™1.292,47""1.531.211.21

SD

0.910.520.980.730.410.47

Motive 5

Score

2.081.662,61"1.741.791.42

SD

0.970.851.130.860.%0.64

Motive 6

Score

3,78"*2.43""1.221.301.301.14

SD

0.580.990.420.620.660.42

Motive 7

Score

3.47""1.791.711.161.051.11

SD

0.650.810.870.370.230.39

**;K.O1. .001.

436 Sakata et al.

{Matsumoto's Flowing Mv.). These results suggest that the raters perceived basicemotions for each motive that were different from other emotions. For each of thesemotives, disgust and fear were rated low, suggesting that Matsumoto's 7 Motivesdo not contain elements that conjure up such emotions.

3.4 Classification by Factor Analysis of Check Lists i and II

After conducting factor analysis for Check List I and Check List II based on fre-quency (selection frequency/number of raters), rotation by the varimax methodwas conducted and factor loading matrix was obtained after the rotation (see Ta-bles 7 and 8).

From Table 7, six factors were extracted from Check List 11 containing 42 terms.The following 8 terms had the absolute value in factor loading matrix of 0.7 orgreater for each emotion factor. For Factor 1, the terms were mechanical, threatening,powerful, aggressive, sharp, rigid, and simple. Out of these 8 terms, 5 also belong to theSharp Mv. in Matsumoto's 7 Motives. Hence, Factor 1 was named as the Sharp Fac-tor. For Factor 2, the 5 terms that had the absolute value in factor loading matrix of0.7 or greater were deep, sad, dark, viscous, and lonely, out of which, 4 terms belong tothe Lonely Mv. in Matsumoto's 7 Motives. Therefore, Factor 2 is named as theLonely Factor. Factor 3 is made up of the 7 terms namely, natural, commonplace, sta-ble, ordinary, weak, calm, and holy. Out of these, 3 terms belong to the Natural Mv. inMatsumoto's 7 Motives. Therefore, Factor 3 is named the Natural Factor. For Factor4, it was the 4 terms, namely, large, majestic, exciting, and imposing, 2 of which belongto Matsumoto's Solemn Mv. Hence, Factor 4 is named as the Solemn Factor. ForFactor 5, it was the 3 terms, namely, elegant, graceful, and flowing. Because all ofthese terms belong to the Flowing Mv. in Matsumoto's 7 Motives, Factor 5 is namedas the Flowing Factor. For Factor 6, only one term merry had the absolute value of0.7 or greater. Because this term belongs to the Dynamic Mv. in Matsumoto's 7 Mo-tives, Factor 6 is named as the Dynamic Factor.

As shown, six factors were extracted by means of factor analysis. When theterms with high factor loading matrix were examined against the categories inMatsumoto's 7 Motives, we found that the terms with the high factor loading ma-trix almost matched Matsumoto's categories for Factors 1,2,3 and 5. However, theSolemn Factor can be said to be different from Matsumoto's Solemn Mv. becausethe term large belongs to the Dynamic Mv. and exciting belongs to the Happy Mv.

We extracted five factors, based on the results of the factor analysis for CheckList I, as shown in Table 8. The terms with the absolute value of factor loading ma-trix of 0.7 or greater were the following four for Factor 1: linear {+), strong {+), vari-able {+), and balanced (-). They were accentuated {-), slow {-i-), constant {+), andspeedy (-) for Factor 2. Similarly, there were the following three terms for Factor 3:diminutive (+), unbalanced {+), and heavy (-I-). For Factor 4, the only term was ex-pansive (+), and for Factor 5, there were two terms, regular (-) and irregular (+).

Each of the five factors just extracted was interpreted for the terms with high fac-tor loading matrix, then we named Factor 1 to be the "Fnergy Factor," Factor 2 to bethe "Time Factor," Factor 3 to be the "Stability Factor," Factor 4 to be the "Expan-

Table 7: Factor Loading Matrix of Check List H

MechanicalThreateningPowerfulAggressiveSharpRigidColdSimpleDeepSadDarkViscousLonelyNaturalCommonplaceStableOrdinaryWeakCalmHolyLargeMajesticExcitingImposingElegantGracefulFlowingMerryBrightDynamicLivelyLight-heartedHappyGleefulHumorousSolemnQuietHeroicUnaffectedSoftTenderWarmEigenvalueVariance {%)

Factor 1Sharp

0.9680.9630.9590.9580.9480.9430.7480.724

-0.078-0.099-0.106-0.106-0.110-0.086-0.109

0.0130.108

-0.098-0.195-0.137-0.069-0.1090.363

-0.117-0.251-0.293-0.409-0.328-0.370-0.480-0.589-0.474-0.557-0.5540.300

-0.142-0.175

0.672-0.170-0.498-0.384-0.67514.64634.870

Factor 1Lonely

-0.171-0.214-^.184-0.182-0.2080.1530.4160.0040.9750.9750.9460.8970.887

-0.104-0.095

0.160-0.1470.3870.2900.371

-0.1500.3410.0820.589

-0.234-0.308-0.390-0.391-0.527-0.509

0.358-0.575-0.580-0.571-0.602

0.5620.674

-0.0400.651

-0.301-0.262-0.25010.70125.479

Factor 3Natural

-0.108-0.052-0.172-0.172-0.188-0.2290.3380.628

-0.1500.064

-0.037-0.122

0.3820.9870.9830.9790.9460.8780.7910.725

-0.090-0.015-0.071-0.058-0.039-0.221-0.076-0.150-0.323-0.343-0.388-0.493-0.492-0.494-0.5340.3990.6590.1290.5120.2280.460

-0.4237.506

17.871

Factor 4Solemn

-0.002-0.0750.0310.038

-0.020-0.0940.207

-0.056-0.0030.0290.2440.375

-0.051-0.060

0.041-0.041-0.201-0.126

0.1380.4980.9240.9200.9050.778

-0.081-0.131-0.153-0.150-0.212-0.146-0.049-0.262-0.270-0.275-0.2640.6650.1340.684

-0.516-0.304-0.390-0.1584.1209.810

Factor 5Flowing

0.1430.1230.1140.1150.1240.1520.3160.2640.1430.1870.1810.1680.2260.027

-0.068-0.059

0.1690.227

-0.4630.250

-0.3020.1390.1650.161

-0.908-0.859-0.7450.239

-0.0310.1240.1570.2840.1060.2060.4320.2350.2250.219

-0.040-0.523-0.4180.0693.3577.992

Factor 6Dynamic

0.0250.0580.0600.0610.0830.0510.1030.0890.0030.0180.0290.0340.0280.0660.0860.0990.0480.0440.1360.086

-0.1400.0770.1030.0660.222

-0.1260.309

-0.799-0.660-0.597-0.590-0.228-0.164-0.0640.0770.0830.1150.1170.1280.4930.4950.5231.6713.977

437

438 Sakataet al.

Table 8: Factor Loading Matrix of Check List I

LinearStrongVariableBalancedAccentuatedSlowConstantSpeedyDiminutiveUnbalancedHeavyExpansiveRegularIrregularCurvaceousSmoothLightsLightEigenvalueVariance (%)

Factor I

Energy

0.9230.9100.821

-0.7410.202

-0.111-0.284

0.2710.038

-0.1110.190

-0.1100.4890,382

-0,680-0,666-fl.672-0.0906.565

36.470

factor 2Time

-0.222-0.091-0.447-0.001-0.9310.8590.857

-0.8050.1960.2430.4420.046

-0.016-0.086

0.2120.368

-0.6510.6015.128

28.490

Factor 3Stability

-0.2400.135

-0.298-0.586-0.255

0.491-0.133-0.202

0.9400.8540.760

-0.057-0.2580.206

-0.287-0.225-0.3330.3932.675

14.860

Factor 4Expansion

-0.0230.0100.0870.2670.1350,0150.2280.463

-0.1370.4310.0190.940

-0.1870.224

-0.630-0.585-0.010

0,2012,060

11.442

Factor 5Regularity

-0.1800.2790.059

-0,166-0.0380.084

-0,3260.0560.108

-0.0660.3720.283

-0.8100.7810.1150.0930.087

-0.6571.1256,252

sion Factor," and Factor 5 to be the "Regularity Factor." The physical movementcomponents of factor scores are listed in Table 9, which shows the intensity of eachcomponent in each of the motives. In this study, the interpreted five factors are thephysical movement components that comprise dance movements.

3.5 Multiple Regression Models of Basic Emotions by MovementComponents

To examine the relation between the perceived basic emotions and physical move-ment factors of the 7 Motives, multiple regression analyses were conducted using

Table 9: Factor Scores of Physical Movement Components

Motive 1Motive 2Motive 3Motive 4Motive 5Motive 6Motive 7

Energy

-0.891-0.016

2.033-0.228

0.409-0.653-0.654

Time

-0.9000.201

-0.6861.1530.952

-1.4050.685

Physical Moi'enient Factors

Stability

-0,1762.045

-0.537-0.4280.263

-0.050-1.116

Expansion

-1,407-0.420-0,3210.7500,4461.600

-0.648

Regularity

-0.550-0.058-0.235-1.736

1.2030.3501,027

Human Body as the Medium 439

the five physical movement factors extracted earlier as explanation variables andthe evaluated scores of the six basic emotions as destination variables. The factorswith the highest significance are shown next as Equations 1 through 6 as the multi-ple regression models.

Happiness - 2.657 - 0.835E** - 0.361 T** - 0.378S** + 0.133R (1)(r2 = 0.995)

Surprise = 1.851 + 0.155E*-0.392T* -0.206S* + 0.051EX + O.n6R (2)(r2 = 0.999)

Loneliness = 2.001 + 0.040E + 0.530T** + 0.515 S** - 0.045 R (3){r2 = 0.998)

Fear = 1.505 + 0.221 E** + 0.095T* + 0.214 S** + 0.051 EX* (4)(r2 = 0.994)

Angry - 1.543 + 0.622E*** - 0.098T* + 0.029 S + 0.041 R (5)(r2 = 0.998)

Disgust = 1.384 + 0.350E* + 0.118S (6)(r2 = 0.940)

*...p < 0.05, * *...p < 0.01, * -̂ * p < 0.001

E..-Energy, T... Time, S...Stability, R...Regularity, EX...Expansion

Further multiple regression analyses were conducted using the evaluated scoresof the basic emotions for the motives, using Equations 1 through 6, as destinationvariables, and using the five physical movement factors of energy, time, stability,expansion, and regularity as explanation variables. As a result, we obtained a mul-tiple regression equation with a high significance for all basic emotions (signifi-cance level p < 0.05). When each emotion was observed, the physical movementfactors that strongly contribute to happiness are energy, time, and stability. Simi-larly, time contributes strongly to the emotion of surprise, whereas time and stabil-ity are the factors of loneliness. The energy and stability factors contribute stronglyfor fear, whereas it was the energy factor for anger and disgust. From these results,it is assumed that the energy, time, and stability physical movement factors contrib-ute to the evaluation of the basic emotions.

3.6 Multiple Regression Models of the 7 Motives Using PhysicaiMovement Factors

Section 3.5 attempted to develop a multiple regression model of the basic emotionsusing physical movement factors as the explanation variables. The next sectiontries to clarify the relation between the physical movement factors and the 7 Mo-tives. First, Check List II made up of 42 terms were classified into seven categories.

440 Sakata et al.

namely, happy, lonely, sharp, natural, solemn, dynamic and flowing. Then the totalscore of the six terms {see Table 3) belonging to each category was obtained. Thiswas then used as the category score. After conducting multiple regression analysesusing the evaluation scores of the five physical movement factors as the explana-tion variables and the category score of the seven categories as the destination vari-ables, the equations with the highest significance are shown next as the multiple re-gression model Equations 7 through 12.

Happy = 48.494 - 32.104E* - 51.085T** - 12.950S (7)(r2 = 0.900)

Lonely - 35.337 + 26.095 T** + 36.969 S** + 3.705EX - 3.399R (8)(r2 - 0.991)

Natural - 29.697 + 24.534 T** + 7.189EX - 20.610R** (9)(r2 - 0.940)

Solemn = 39.846 + 30.440 T* + 31.743 S* (10){r2 = 0.755)

Sharp = 37.967 + 71.356 E* - 19.023 T - 12.798 S - 9.439EX (11)(r2 = 0.927)

Dynamic = 47.741-23.491 T* + 10.178R (12)(r2 = 0.672)

*...p < 0.05, * \..p < 0.01, ***p< 0.001

E...Energy, T... Time, S...Stability, R...Regularity, EX...Expansion

As shown by Equations 7 through 12, a regression equation with high signifi-cance was obtained for all of the motives, except for the flowing motive (signifi-cance level p < 0.05). From the values of the coefficient concerned with each motive,we can find that the physical movement factors of Hme and energy contributestrongly to the happy motive, while it is stability and time for the lonely motive,time, and regularity for the natural motive, time and stability for the solemn mo-tive, energy for the sharp motive, and time for the dynamic motive.

For the flowing motive, it was not possible to obtain an effective regressionmodel (with significance level of 0.05 or less).

4. DISCUSSION

This study used Matsumoto's 7 Motives newly performed by a dancer as the stimu-lus to conduct an evaluation experiment and examined the kind of images andemotions perceived by the raters. Then researchers attempted to develop a multi-ple regression model that would show the effects of physical movement factors onthe impressions and emotions.

Human Body as the Medium 441

4.1 Emotional Information in Body Movements

It was suggested that the raters were able to perceive impressions particular to eachmotive. At the same time, basic emotions, except the emotions of disgust and fear,were roused. Therefore, we can say that it was confirmed that Matsumoto's 7keywords, namely, happy, lonely, sharp, natural, solemn, dynamic, and flowing inthe emotion value of dance movements, and the happiness, surprise, loneliness,and anger emotions, out of the six basic emotions by Ekman were perceived by thesame body movements, that is, dance movements. So the raters were considered tohave perceived the images and emotions particular to the different motions corre-sponding to the moven:ient of the performer.

Ekman and Friesen (1967) stated, "... the location of the body trunk does not of-fer any information regarding the quality of emotion. Rather, it can only give infor-mation on a larger emotional status than particular emotion, if any." In this way, hehas a negative view about the information offered by the body {Ekman & Friesen,1975). In the video-tape "Quality of Movement and Emotional Value<Matsumoto>, Japan Association of Physical Education for Women," which has arecording of the 7 Motives {Matsumoto, 1985) and was used for this study, the per-former shows no facial expression throughout. Patterned after this video, the vi-sual stimuli used in this study also show the performance with a complete lack offacial expressions. The raters were told in advance to "rate the performance basedonly on body movements since the performer would not show any facial expres-sions." TTie results show it is possible to perceive emotional information from thebody-mediated information without facial expressions. Thus, they suggest that thebody does not only have an additional or secondary function, but also plays an im-portant role as an independent medium in human communications.

4.2 Evaiuation of Disgust and Fear

As described earlier, this study was not able to find a dance movement that con-jured up fear and disgust.

On the other hand, gestures contain typical movements that express fear anddisgust (e.g., Ekman & Friesen, 1975; Morris, 1999). However, the participants ofthis study were dance movements and not gestures or facia! expressions. This maysuggest that dance movements are a different medium from gestures and facial ex-pressions in the expression of fear and disgust^ As described earlier, consideringthat more than one basic emotion was perceived from one motive, the seven emo-tional qualities, which are said to be typical in dance, are suggested to be differentin representation from the six basic emotions, which are considered typical in facialexpressions.

^We continue to study the physical movements that represent the basic emotions (Ekman & Friesen,1975) in dance movements. Here again, difficulties are suggested for the creation of dance movementsthat express disgust (Shiba, Kikuchi, & Odaka, 1999-2001).

442 Sakata et al.

4.3 Physical Movement Factors and Basic Emotions

We extracted five elements as clues for recognizing the impressions and emotionsfrom each motive and determined these five elements, namely, energy, time, stabil-ity, expansion, and regularity, to be the physical movement factors comprising the 7Motives. In the multiple regression analyses we conducted using the five physicalmovement factors as perceived emotions as explanation and destination variablesrespectively, significant regression equations were obtained for each motive. Out ofthe five physical movement factors, one of energy, time, and stability factors isgreatly involved in the evaluation of each emotion, although involvement was lowfor expansion and regularity factors. Therefore, it can be assumed that the physicalmovement factors affecting the evaluation of the basic emotions perceived fromdance movements are energy, time, and stability.

Matsumoto (1985) determined that energy, time, and design were the three fac-tors comprising a physical movement, and based on these factors, she formulatedCheck List 1 made up of 18 terms. Furthermore, Matsumoto used these three factorsof physical movements as the foundation to create the movements of the 7 Motivesused as the stimulus in this study. The results of this study show that the energyand time factors, out of the three physical movement factors of energy, time, andstability that are suggested to be involved in the recognition of the basic emotionsin dance movements, are the same as Matsumoto's physical movement factors. Theresults also indicate that "stability" can be considered a comprising factor of "spa-tial design." Therefore, it is assumed that the dance movements of the 7 Motivescreated on the basis of time, energy, and spatial design functioned as indexes to al-low the raters to differentially recognize the basic eniotions.

4.4 Physical Movement Factors and Matsumoto's 7 Motives (EmotionalQuality)

The multiple regression analyses conducted to examine the relations between thefive physical movement factors and the seven keywords—happy, lonely, sharp,natural, solemn, dynamic, and flowing—we found that either of the four physicalmovement factors, namely, energy, time, stability, and regularity, were heavily in-volved in the evaluation of all the motives, except for the flowing impression.Therefore, it is considered that the energy, time, and stability factors, which equallyaffect the basic emotions, as well as regularity, are the physical movement factorsthat affect the evaluation of the quality of emotions.

A multiple regression model for the flowing impression could not be obtained.However, because the flowing factor was extracted in the Check List II factor analy-ses, the raters seemed to have surely recognized the flowering factor. Nevertheless,the failure to develop a multiple regression model between the terms belonging tothe word group of flowing impression and the physical movement factors can beattributed to the fact that the terms such as soft, tender, flowing, elegant, graceful, andzvarm also possessed characteristics of physical movement factors, or a type ofmovement, instead of a quality of emotion that can be defined as "fluidity." Shiba et

Human Body as the Medium 443

al. (2001) pointed out the possibility that the terms which indicate the quality ofemotion and type of n:iovement exist intermixed in Check List II, comprised of 42terms, and Check List I, comprised of 18 terms. Therefore, it will be necessary in thefuture to reexamine the grouping method used for Check List II, which is con-cerned with the quality of emotion, and the grouping method used for Check ListII, which is concerned with movement types, based on time, energy, and spatial de-sign factors.

5. CONCLUSIONS

To clarify mechanisms involved in the human recognition of body-mediated infor-mation, we examined what kinds of impressions and emotions are perceived frombody movements in dance with typical body movements directly connected toemotions. Using Matsumoto's 7 Motives, this study attempted to develop multipleregression models to investigate the effect of physical movement factors on impres-sions and emotions conveyed through the body-mediated information.

The items we have clarified in this study are summarized next.

1. Because characteristic emotional in:ipressions were recognized for each of thedance movements in Matsumoto's 7 Motives, the human body, excluding fa-cial expressions, works as independent media for communicating emotions.

2. Matsumoto's 7 Motives do not include elements that conjure up disgust andfear.

3. Five factors—namely, energy, time, stability, expansion, and regularity-were extracted from Matsumoto's Check List I, and six factors, namely,sharp, lonely, natural, flowing, dynamic, and solemn were extracted fromMatsumoto's Check List I.

4. The emotional impressions perceived from each motive were greatly af-fected by the physical movement factors, namely, energy, time, and stability,whereas the impact of these factors varies for each impressed and expressedemotion.

5. The grouping method of the quality of emotion (Check List I) and move-ment type (Check List I) requires further reexamination.

The re-creation or representation of human movements by robots and comput-ers requires the identification of human models of body-mediated information as aprerequisite. In this sense, clarification of any regularity concerning dance move-ments will set the basis of human interface. In an age where the importance of non-verbal communication increases constantly, research on dance is expected to play agreat role in human interface research.

By means of three-dimensional motion analyses, Sakata et al. (2001) extractedthe elements of body movements that cause different impressions. It is hoped thatthe relation between the physically measured values obtained in that study and thephysical movement factors extracted in this study will be examined further. In ad-dition, it is necessary to conduct further research on usual physical movements, not

444 Sakataetal.

only dance movements. Moreover, it is necessary to examine terms that can ade-quately evaluate physical movements in addition to the basic movements and theterms contained in Matsumoto's Check Lists. At present, we are in the process ofcreating a commonly usable scale to measure emotional values of different media,while examining the relations between factor structure particular to different me-dia types, through the projects named "Basic Study on Correlation between Differ-ent Emotional Media (Shiba, Kikuchi, & Odaka, 1999-2001)" and "Verification ofBasic Emotions Common to Different Emotional Expressions and Comparison Be-tween Different Cultures (2003)."

REFERENCES

Ekman, P., iSc Friesen, W. V. (1967). Head and body cues in the judgment of emotion; A refor-mulation. Perceptual and Motor Skills, 24, 711-724.

Ekman, P., & Friesen, W. V. (iy75). Unmasking the face. Englewood Cliffs, NJ: Prcnticc-Hail.Imagawa, K., Taniguchi, R., Arita, D., Matsuo, H., Lu, S., & Igi, S. (2000). Recognition of local

features for camera-based sign-language recognition system. The journal of the Institute ofImage Information and Teleinsion Engineers, 54, 848-857.

Kurokawa, T (1992). Gesture coding and a gesture dictionary for a nonverbal interface.lEICE Trance. Fundamentals, E75-A, 112-121.

Kurokawa, T., Moriuchi, T, & Watanabe, H. (1993). Bidirectional translation between signlanguage and Japanese for communication with deaf-mute people. Advances in HumanFactors/Ergonomics, 19B, 1109-1114.

Lim, S., & Tamura, S. (1999). Analysis of American Sign Language sentence by signing spaceand directional verb. Journal of Human Interface Societ}/, 1,1-6.

Matsumoto, C. (1968). An introduction of kinesiology (pp. 276-283). Tokyo: Taishukan Inc.Matsumoto, C. (1985).Dance research: Problem situation and learning of problem solving

II—Qualities of movements and feeling values. Lecture at the Tenth Congress of the Interna-tional Association of Physical Education and Sport for Girls and Women, Lerwick.

Morris, D. (1994). Bodytatk—A world guide to gestures. London: Jonathan Cape.Sakata, M., Shiba, M., & Iwadate, Y. (2000). Body movement as KANSEI information—Emo-

tional expression of figure skaters immediately after their performances. Bulletin of theFaculty of Human Development, 8, 231-240.

Sakata, M., Uchida, S., Hama, A., Shirakawa, T., & Maeda, M. (2001). A three-dimensionalanalysis of dance movement, journal of Physical Education and Sports Science, 10, 49-57.

Shiba, M., Kikuchi, M., & Odaka, N. (1999-2001). Basic study on correlation between differ-ent emotional media. The Reports of the Consigned Research by ATR Media Integration &Communication Research Laboratories. Kyoto, Japan: ATR.

Shiba, M., Sakata, M., & Tsubokura, K. (2001) The expressive character of the dance move-ments based on the expression word of the face—The relation of the facial expression andbody expression. Bulletin of the Faculty of Human Development, 8, 269-282.

Shiba, M., Uozumi, K., Kikuchi, M., et al. (2003). Verification of basic emotions common todifferent emotional expressions and comparison between different cultures. FundamentalResearch in Grant-in-Aid for Scientific Research (C)(2)(126W08V, 1-93.

Wundt, W. (1922). Grundriss der psychologic. Leipzig: Kroner.Yoshida, A., Renge, K., Kaga wa, C , Sako, H., & Maiya, K. (1989). A handbook of behavioral sci-

ence. Fukumura Inc.